Faculty of Social and Behavioural Sciences
Graduate School of Childhood Development and Education
Measuring state symptoms of PTSD in
children and the relationship between
PTSD symptoms and heart rate
Research Master Child Development & Educational Sciences Thesis 2
I. Aarts (6077765)
Dr. L. Van Rijn-Van Gelderen – Prof. Dr. G. J. Overbeek 07-07-2016
Abstract
State symptoms of posttraumatic stress disorder (PSTD) and the association between state and trait symptoms of PTSD, and heart rate measures have not yet been studied in children. In the current study, 74 children aged 8-18 participated, all with a clinical diagnosis of PTSD. Children participated in a script driven imagery paradigm during which their heart rate responses were measured. The Response to Script Driven Imagery Scale (RSDI) was used to assess state symptoms of PTSD. In contrast to in adults, almost all items correlated with each other and only two factors were found. It is unclear why the factor structure is different in children. Possibly PTSD is less differentiated in children than adults, or children are not able to report on state symptoms as well as adults. The two factors of the RSDI correlated
positively with heart rate reactivity but not heart rate variability reactivity. Unexpectedly, trait measures of PTSD did not correlate with heart rate and heart rate variability reactivity.
Exposure to a traumatic event can lead to posttraumatic stress disorder (PTSD). Criteria for PTSD are reexperiencing, avoidance and heightened arousal (American
Psychiatric Association [APA], 2000). PTSD is commonly diagnosed with a semi-structured interview to assess (trait) symptoms of PTSD, for example the Clinician Administered PTSD Scale (CAPS; Blake et al., 1995).
Reporting trait-symptoms, however, might be hindered by bias associated with reconstructive recall and can therefore lead to poor diagnostic discrimination (Hopper, Frewen, Sack, Lanius & van der Kolk, 2007). It might thus be more useful to assess state symptoms of PTSD rather than trait symptoms. Self-report of state symptoms might
complement current diagnostic procedures that focus mainly on assessing trait symptoms, and thereby help improve diagnostic discrimination (Hopper, Frewen, Sack et al., 2007). The Response to Script Driven Imagery Scale (RSDI) was developed to measure state symptoms of PTSD in response to a trauma-related stimulus (Hopper, Frewen, Sack, et al., 2007) and has been studied in three samples of adults. Results indicated a factor structure with three factors that held across the samples. These factors were reexperiencing, avoidance and dissociation. Also, there was evidence that supported convergent and discriminative validity. Later research has shown different neural correlates for these three dimensions (Hopper, Frewen, van der Kolk & Lanius, 2007), providing empirical support for the three different state PTSD response types measured with the RSDI. The RSDI has thus far only been studied in adults.
Findings in PTSD research with adults cannot simply be generalized to children, for several reasons. Firstly, some symptoms of PTSD as described in the Diagnostic and Statistical Manual of Mental Disorders (4th ed., text rev.; DSM-IV-TR; APA, 2000) can express themselves differently in children and adults. In adults, ‘recurrent distressing dreams
of the event’ is a symptom, while in children these dreams do not need to be related to the traumatic event. Also, ‘acting or feeling as if the traumatic event were recurring’ can involve trauma-specific reenactment in children. Secondly, there is a discrepancy in lifetime
prevalence rates of PTSD in children and adults. Prevalence rates of 0.5% have been found in children, compared to 8% in adults (Fairbank & Fairbank, 2009), whereas the rates of
experiencing a traumatic event are already quite high in childhood. Copeland, Keeler, Angold and Costello (2007) found that by the age of 16, 67.8% of children experienced at least one potentially traumatic event. One possible explanation for this large difference in prevalence of PTSD is that the DSM symptoms are insufficiently developmentally informed (Fairbank & Fairbank, 2009). This implies that research done in adults cannot be readily generalized to children. Finally, research indicates that the severity of a potentially traumatic event predicts the development of PTSD differently in children and adults. In children the subjective
experience of an event predicts the development of PTSD-symptoms better than the objective level of severity of the event (Verlinden et al., 2013). In adults, on the other hand, the
subjective experience of the event does not add anything to the prediction of PTSD
development over the objective level of severity (Verlinden et al., 2013). Because of these differences between children and adults and the current unavailability of a self-report measure to assess state PTSD symptoms in children, one of the goals of the current study is to test the validity of an age-adapted version of the RSDI in children and adolescents.
To complement the study of the validity, we will study the relationship between self-reported state symptoms and psychophysiological measures. Hopper, Frewen, Sack et al., (2007) used heart rate reactivity as a measure for convergent validity. Heart rate reactivity is a psychophysiological measure of the change in heart rate after a cue. It is often used in
2007). Because psychophysiological measures do not rely on the accuracy of self-report and observations of the clinician (Orr & Roth, 2000) they could complement subjective measures as an objective (outcome) measure. Initially, psychophysiological measures were used to assess biological correlates of the DSM-criterion ‘physiological reactivity on exposure to internal or external cues that symbolize or resemble an aspect of the traumatic event’
(Criterion B.5 in the DSM-IV-TR; APA, 2000). Psychophysiological measures can be used to assess both state and trait changes in physiological reactions. With respect to state PTSD measures, there was some evidence for a connection between the state symptoms and heart rate reactivity in the study of Hopper, Frewen, Sack et al., (2007). They found a positive correlation between the RSDI reexperiencing scores, and heart rate reactivity. In one of their three samples, they found a small negative correlation between RSDI Dissociation and heart rate reactivity. With regard to trait symptoms, in adults, the dissociative subtype is
characterized by emotional overmodulation, which means that there is too much inhibition of the emotional regulation system. In the reexperiencing/hyperarousal type, there is too little modulation of the affective response to traumatic memory (Lanius et al., 2010).
There are also other psychophysiological measures that are used in the PTSD research. Heart rate fluctuates influenced by the autonomic nerve system. These changes in heart rate over time are named heart rate variability (Tan, Dao, Farmer, Sutherland & Gevirtz, 2011). A negative relationship between low frequency heart rate variability and state PTSD symptoms has been found in a recent study (Green et al., 2016), but no distinction between different symptoms was made.
Trait-like changes in physiology include heightened basal heart rate in PTSD patients as compared to non-trauma exposed persons (Buckley & Kaloupek, 2001). In his meta-analysis on psychophysiological measures of PTSD in adults, Pole (2007) found that PTSD
was in general related to higher resting heart rate, skin conductance and diastolic blood pressure compared to both healthy control subjects and trauma exposed control subjects.
To establish state physiological changes in PTSD, the script-driven imagery (SDI) paradigm is widely used. In the SDI-method, subjects are exposed to a trauma script, a narrative of their own traumatic experience. Their psychophysiological responses are then measured (Pitman et al., 1990). Pole (2007) found in his meta-analysis that in reaction to idiographic trauma cues (e.g. a trauma script) persons with PTSD showed larger responses for facial muscles, heart rate, skin conductance, and diastolic blood pressure compared to people without PTSD.
The findings in adults are quite robust, but there has been less research in children. According to Kirsch, Wilhelm and Goldbeck (2011), physiological findings in adults cannot simply be generalized to children. The developmental life trajectory of children should be taken into account, for example because conflicting results have been found between pediatric and adult PTSD with respect to physiology.
There has been some research with regard to psychophysiology in PTSD in children. With respect to trait measures, children exposed to marital violence had significantly higher heart rates at the start of the experiment than a clinical comparison group consisting of children with mental health problems such as anxiety and disruptive behavior, but with no experience of marital violence (Saltzman, Holden, & Holahan, 2005).
Scheeringa, Zeanah, Myers and Putnam (2004) studied responsivity to trauma cues in children under the age of seven, thus assessing state characteristics of PTSD. They asked children to recall the memory of a traumatic event and asked questions about this event. Traumatized children reacted to the trauma cues with decreased heart rate period (comparable to heart rate variability) in comparison to non-traumatized control children. In another study,
the heart rate of exposed children was higher than that of controls after an interview about their traumatic experiences (Saltzman, Holden & Holahan, 2005). This difference increased from pre-interview to post-interview. On the other hand, another study found no such increase in heart rate (Jones-Alexander, Blanchard & Hickling, 2005). They compared two groups of children that had been in an accident, of which one group developed PTSD and one did not develop PTSD, and a control group who had not been in an accident. The PTSD group did report higher subjective distress than the other two groups after hearing their trauma script, but there was no difference in heart rate response to trauma script between the groups.
In short, the results with regard to physiology are heterogeneous in children. Some studies found higher heart rate after a trauma script (Saltzman et al, 2005; Scheeringa et al, 2004), while another study found no difference (Jones-Alexander et al., 2005). In adults, on the other hand, the finding that heart rate goes up after hearing a trauma script is quite robust (e.g. Pole, 2007). Because of these heterogeneous findings and the idea that results from adult studies cannot be generalized to children, the second part of this study will focus on the relationship between both trait and state symptoms of PTSD (reexperiencing, avoidance and dissociation), and heart rate data.
The Present Study
The first research question is whether the RSDI can be used to study self-report of state symptoms of PTSD in children. We hypothesized that the three factor structure of the RSDI would hold in children, with reexperiencing, avoidance and hyperarousal as the three factors. The second research question focuses on the relationship between trait and state measures of PTSD in children and heart rate data. For both state and trait measures, we expected the same findings as in research on adults (Hopper, Frewen, Sack et al., 2007). So, we hypothesized that there would be a positive relationship between heart rate and
reexperiencing a traumatic event, and that there would be a negative relationship between heart rate and dissociation. This would mean that a higher heart rate response would be associated with more intense reexperiencing of traumatic events, and that a lower heart rate response would be associated with higher dissociation. With regard to avoidance, we did not put forward a specific hypothesis. It would be possible that children avoid the script by thinking about something else – not showing a heightened physiological response – but it would also be possible that avoiding the script causes stress, therefore heightening the
physiological response. No specific hypothesis was made for heart rate variability, because no earlier study has looked at the relationship between heart rate variability and specific state symptoms of PTSD.
Method
Participants
The participants were 74 children (44 girls) aged 8 – 18 (Mage = 12.71, SD = 3.24)
recruited at the Centre for Trauma and Family at De Bascule, the department of child and adolescent psychiatry of the Academic Medical Centre in The Netherlands. All children experienced one or more traumatic events and were referred to the centre for treatment. All children met the criteria for posttraumatic stress disorder (PTSD) or partial PTSD. We also included children with partial PTSD, because PTSD symptoms in this group also cause clinically meaningful functional impairment, despite the fact that the level of impairment is often less severe compared with full PTSD (Stein, Walker, Hazen & Forde, 1997).
We excluded participants who met the following criteria: acute suicidality, pregnancy, known presence of a significant medical illness (including cardiac arrhythmias), or meeting the criteria of one or more of the following DSM IV diagnoses: a psychotic disorder,
substance use disorder, a pervasive developmental disorder, mental retardation or the use of medication that influences the central nervous system or the HPA axis.
Procedure
All children referred to the Centre for Trauma and Family were asked to participate in a study that aimed to compare the treatment efficacy of two trauma treatments, eye movement desensitization and reprocessing (EMDR) and trauma-focused cognitive behavioural therapy (TF-CBT) and biological correlates of PTSD. The current study is part of this larger study. Children were measured at baseline, before the start of the treatment. As part of the diagnostic procedure, all children above eight years old were interviewed with a semi-structured
interview. Parents were also interviewed, and both parents and children filled out several questionnaires. If children met inclusion criteria and both child and caregiver gave informed consent, they came to the Centre again for a physiology appointment and an MRI-scan. Both appointments took about one hour. Children received a gift card (15 euro) for their
participation. The study was approved by the medics ethics comity of the Academic Medical Centre (NL35971.018.11).
When children came to the centre for a physiology appointment, physiological and endocrinal measures were taken during a script driven imagery protocol (SDI).All
participants were seated in front of a laptop in a temperature controlled room (20 degrees Celsius). The examiner was seated approximately 1 meter behind the participant throughout the experiment. The participants were connected to a portable physiological monitoring device which provided continued monitoring throughout the remaining procedure. Comparable to earlier studies at De Bascule (e.g. Jonkman et al., 2013), electrodes were placed by the examiner (five on the chest and two on the back) and signals were checked at
the start of the recording. Also, two electrodes were attached to the fingers of the children, to measure skin conductance. These electrodes were connected to a VU-AMS device (VU University Ambulatory Monitoring System; De Geus, Willemsen, Klaver & van Doornen, 1995). Firstly, there was a baseline phase of 30 seconds in which children watched nature videos. Secondly, there was a 1 minute neutral script phase (30 seconds neutral script about brushing your teeth, 30 seconds imagery in which children were instructed to sit with their eyes closed and think about the script they just heard), followed by a 59 to 61 seconds trauma script phase. In this phase, children heard a script of 29-31 seconds of their most disturbing traumatic event. This script was told in the present time and included children’s thoughts and feelings at the moment of the traumatic experience. Afterwards, there was a 30 second imagery period. During the imagery periods, children had to sit with their eyes closed and think about the story they just heard. Finally, the children filled out the Responses to Script Driven Imagery scale (RSDI, Hopper, Frewen, Sack et al., 2007) while still attached to the VU-AMS. After each phase (baseline, neutral and trauma), children were asked to rate their emotional valence and tension. At the end of the SDI procedure participants were asked to complete two questionnaires.
Measures
PTSD was measured with a semi-structured trauma interview, the Dutch version of the Clinician-Administered PTSD Scale for Children and Adolescents (CAPS-CA; Nader et al., 2004). The interview was conducted by a trained researcher. The CAPS-CA was used to diagnose PTSD and also provided an overall severity score for PTSD. The CAPS-CA assesses all possible symptoms of PTSD according to the DSM-IV-TR. For every symptom, the
4 = four or more times a week in the past month). Intensity was also rated from 0 to 4 (0 = not a problem, none; 4 = a whole lot, extreme, incapacitating distress, cannot dismiss memories, unable to continue activities). Clinicians could adjust this score according to their judgement, for example if they suspected that a child was very avoidant. So, the child rated each
symptom and the clinician could adjust this score. A child was diagnosed with PTSD when there was at least one symptom of reexperiencing, at least three of avoidance and at least two of heightened arousal. The (clinician corrected) scoring rule of frequency score ≥ 1 and intensity score ≥ 2 was used (Weathers, Ruscio, & Keane, 1999). Also, to receive a diagnosis of PTSD a child had to report a suffering score of 2 or higher rated on a 4-point scale (0 = not a problem, none; 4 = a whole lot, extreme, incapacitating distress, cannot dismiss memories, unable to continue activities) on at least one of four areas; general suffering, social problems, problems in school and regressive symptoms (such as wetting the bed). Partial PTSD was diagnosed when children fulfilled two of the three criteria completely, or had at least one symptom of each criterium. The CAPS-CA has been shown to have good psychometric properties in a Dutch sample of 112 children aged 8 to 18, with Cronbach’s α of 0.83 for the total scale, and interrater reliability ICC 0.97-0.99 (Diehle, de Roos, Boer, & Lindauer, 2013).
To assess trait symptoms of PTSD, total CAPS-CA scores of reexperiencing,
avoidance (only those about effortful avoidance, see King, Leskin, King, & Weathers, 1998) and dissociation (see for example Wolf et al., 2012) were used. Internal consistencies
(Cronbach’s α) were calculated for these subscales of the CAPS-CA: reexperiencing (N = 70,
α = .91), effortful avoidance (N = 70, α = .67; only based on two items) and dissociation (N =
60, α = .77).
To complement the diagnosis based on children’s report, the PTSD-scale of the
was used. There is no research about the psychometric properties of the PTSD module of the ADIS-P. Parents/caregivers rate every symptom of PTSD as either absent or present. The same criteria were used to diagnose PTSD or partial PTSD as with the CAPS-CA.
Furthermore, a suffering score of at least 4 was needed to fulfil a diagnosis of (partial) PTSD on a scale of 0 to 8 (0 = no impairment, 8 = a lot of impairment). ADIS-P scores were only used for determining the diagnosis of PTSD, not for further analyses.
The RSDI (Hopper, Frewen, Sack et al., 2007) was completed by children directly after the trauma imagery period in the SDI protocol. It consists of 11 items that correspond to three factors: reexperiencing, avoidance and dissociation. Example items are “did you feel as though the event was reoccurring, like you were reliving it?” (reexperiencing), “did you avoid thoughts about the event?” (avoidance) and “did you feel like you were in a fog?”
(dissociation). Children rate answers to these questions on a scale of 0 (not at all) to 6 (a great deal). The original version of the RSDI was adapted to better suit children’s understanding and translated into Dutch. The adapted Dutch translation was back-translated and approved by the original author. Cronbach’s α for the total RSDI scale was .873 in the current study.
Heart rate (HR) and heart rate variability (HRV) were calculated with VU-AMS software version 3.9 to (pre-) process the data. After visual inspection and manual correction for artefacts, interbeat interval time series and respiration signals were derived from the ECG and ICG signals. From these signals mean heart rate and respiratory sinus arrhytmia (RSA) were computed for each of the three conditions (baseline, neutral imagery and trauma imagery). RSA is a measure of heart rate variability (HRV) data. Heart rate was directly derived from the interbeat interval time series. RSA was extracted combining the interbeat interval time series and respiration signals using the peak-valley method (De Geus et al., 1995). In short, this consists of subtracting the shortest interbeat interval during heart rate
acceleration during inspiration from the longest interbeat interval during heart rate deceleration during expiration of every breath..
Data Analysis
To assess whether the RSDI can be used in children, a confirmatory factor analysis was conducted on the RSDI data. The model was compared to that of Hopper, Frewen, Sack et al., (2007), with three latent variables (reexperiencing, avoidance and dissociation). Items 1-4 of the RSDI were expected to load on reexperiencing, 5-7 on avoidance and 8-11 on dissociation.
Data were analyzed with R (version 3.2.3) with the package Lavaan (version 0.5-20). A Confirmatory Factor Analysis (CFA) was used to evaluate the three-factor structure from Hopper, Frewen, Sack et al., (2007). Model fit was evaluated with the Comparative Fit Index (CFI), with a CFI ≥ .95 indicative of good fit and the root mean square error of approximation (RMSEA) < .08.
When a non-fitting model occurred, exploratory factor analysis (EFA) was conducted in SPSS (version 19).It was assumed that the factors would correlate; therefore an oblique rotation was used. To assess the optimal number of factors, the following criteria were used: a scree plot, eigenvalues larger than one and a parallel analysis (Tabachnik & Fidell, 2013). In parallel analysis eigenvalues are extracted from random data sets that are comparable to the actual data set in terms of number of cases and variables. Factors are retained if eigenvalues from the random data set are smaller compared to eigenvalues from the actual data set (O’Connor, 2000). For the parallel analysis, the method and SPSS syntax described by O’Connor (2000) were used. Factor loadings larger than |.40| are considered fair and were retained (Tabachnik & Fidell, 2013).
For the heart rate data, Q-Q plots were made to check for normality of the data. The HRV data was log transformed. Then, paired t-tests were run on the HR and HRV data, to test whether heart rate changed from neutral imagery to the trauma imagery period. A reactivity score was calculated by subtracting the HR during neutral imagery from HR during trauma imagery. The same was done for HRV. Then, Pearson correlations were made between the scores on the CAPS-CA subscales (reexperiencing, avoidance, dissociation), the RSDI factors found in the study and the HR and HRV reactivity scores.
Results
General results
First, we conducted exploratory analyses. For severity of PTSD (M = 52.17, SD = 24.28) there was a significant difference between boys and girls (t(69) = -5.33, p < .001, d = 1.24), with girls having on average higher PTSD severity scores (Mboys = 36.75, SD = 16.79, Mgirls =
61.98, SD = 23.30). Also, children aged 8-12 (n = 33, M = 40.45, SD = 21.34) scored lower than children aged 13-18 (n = 39, M = 62.08, SD = 22.31), t(70) = -4.180, p < .001, d = 0.99.
The Responses to Script Driven Imagery scale in children
Normally distributed responses to the RSDI were confirmed with Q-Q plots. Means, standard deviations and correlations of the RSDI have been calculated, see Table 1. The correlations showed that almost all items of the RSDI are significantly positively correlated to other items, with a few exceptions. Item 10 is only significantly associated with items 2, 3, 4 and 11. Item 11 is not significantly related to item 5 and 7 but is positively related to all other items.
For the confirmatory factor analysis (CFA), the three-factor model was built to analyze the data. A correlation > 1 appeared between the factors Reexperiencing and Avoidance. Because of this non-admissible value in the solution, model fit measures and model parameters are not reported for the CFA.
Next, an exploratory factor analysis (EFA) was run on the data. Two eigenvalues were larger than 1 (5.135 and 1.609 respectively; the eigenvalue for the third factor was 0.883), indicating that extracting two factors would provide the best results. The results of the parallel analysis also indicated that two factors would be the optimal solution with the first three generated eigenvalues being 1.683, 1.472 and 1.316.
Together, the two factors explained 61.30% of the variance (46.68% of the total by the first factor). In Table 2, the factor pattern and factor structure matrices are reported. Also, the initial and extraction communalities are given. There was a medium sized correlation between the two extracted factors, r = .430. All items except item 10 and 11 loaded on the first factor. Items 2, 3, 4, 6, 10 and 11 loaded on the second factor. These are three reexperiencing items (2-4), an avoidance item (6) and two dissociation items (10 and 11).
Relationship between PTSD symptoms, heart rate and heart rate variability
After logtransforming the HRV data, the Q-Q plots indicated that the data were normally distributed. See Table 3 for the descriptives of the trait reexperiencing, avoidance and dissociation scores, HR and HRV during neutral imagery and trauma imagery.
From neutral imagery to trauma imagery, paired-samples t-tests for differences revealed a significant increase in heart rate from neutral to trauma imagery, t(73) = -2.418, p = .018, d = -0.286. This means that heart rate was higher (M = 81.29) in the trauma imagery phase than in the neutral imagery phase (M = 79.59). There was no difference in HRV from neutral to trauma imagery, t(71) = 1.607, p = .113, d = 0.183.
Correlation analysis revealed no significant correlations between both HR and HRV reactivity scores and reexperiencing (r = .198 and r = -.070 respectively), avoidance (r = .182,
r = -.024) and dissociation (r = .118, r = .172), see Table 4.
Significant positive correlations were found for the first RSDI factor with trait reexperiencing, avoidance and dissociation. The second RSDI factor correlated positively with reexperiencing and avoidance. There was a significant positive correlation between HR reactivity and both RSDI factors (r = .278 and r = .243, respectively). The RSDI factors did not correlate significantly with HRV reactivity (r = -.080, r = .058).
Discussion
State measures of posttraumatic stress disorder (PTSD) have been used to complement diagnostic procedures in adults, but not yet in children. The results of the current study
indicated that the three factor structure of the original Responses to Script Driven Imagery scale (RSDI) does not hold in children. Exploratory factor analysis lead to two factors, with all but two items loading on the first factor and multiple items loading on both factors. The three factors of the original RSDI obtained in adults were not recognizable in these results in a population of children aged 8 to 18. Trait measures of PTSD did not correlate with heart rate (HR) and heart rate variability (HRV) reactivity scores. Both RSDI factors showed a positive association with HR reactivity, but no association with HRV reactivity. Both RSDI factors showed a positive relationship with trait reexperiencing and avoidance, but only the first RSDI factor was positively associated with trait dissociation.
RSDI in children
Our findings in children with PTSD were unexpected and do not correspond to previous findings in adults with PTSD, where the three factor structure of the RSDI was confirmed
(Hopper, Frewen, Sack et al., 2007). The results from the correlation analysis showed
correlations in the expected direction, but also show unexpected correlations between items of different scales of the original RSDI. Almost all items of the RSDI correlated with each other in the current study, indicating that higher scores on one state symptom are related to higher scores on other symptoms. This could imply that PTSD is less differentiated in childhood than in adults. This might help explain the large difference in prevalence rates of PTSD between children and adults, as described in the introduction. If PTSD manifests itself less
differentiated in children than in adults, diagnostic procedures might have to be reconsidered and changed for children. This is in line with Meiser-Stedman, Smith, Glucksman, Yule and Dalgleish (2008) who showed empirical support for a different diagnostical algorithm for children. It would be interesting to repeat the analysis of the RSDI for younger and older children separately to provide support for this idea. If symptoms of PTSD differentiate more with development, a factor structure more comparable to adults than the one found in the current sample would be expected in older children. The study by Contractor et al., (2013) shows a difference in factor structure of PTSD symptoms between children of different ages, supporting the notion that age plays a role in the manifestation of PTSD symptoms.
A first other possible explanation for the findings is that children are not able to report adequately on their own state symptoms of PTSD. Trait severity scores of PSTD differed in this study for younger and older children, indicating a difference in how children experience or report on their own symptoms. State symptoms of PTSD in children have not been studied before, therefore the current result cannot be compared to other studies. However, there are some self-report measures of state anxiety that have shown good psychometric properties in children of 8 years and older, like the State-Trait Anxiety Inventory for Children (STAI-C; Spielberger, 1973). The STAI-C consists of two scales, one for trait anxiety and one for state
anxiety. So, children are able to report on at least some state symptoms. In the STAI-C, however, no distinction is made between different forms of state symptoms, like in the RSDI.
In the validation study of the interview used to report trait measures of PTSD, the Clinician-Administered PTSD-Scale for Children and Adolescents (CAPS-CA; Diehle, de Roos, Boer & Lindauer, 2013), lower Cronbach’s αs were found for the age group 8-12 than in the group of children aged 13-18. Also, most broad mental health self-report measures are meant for children of 12 years and older and not for younger children (e.g. Deighton et al., 2014). Again, it would be interesting to study if there are differences between age groups in whether the RSDI can be used to assess state symptoms. Unfortunately, the current group was too small to make such comparisons. Future studies should address potential differences between how well younger and older children are able to report on state symptoms.
It is also possible that PTSD symptoms are fundamentally different from anxiety symptoms and even though children can report on some state symptoms (such as measured with the STAI-C) they cannot report well on PTSD symptoms. For instance, one of the symptoms for PTSD is a restricted range of affect (C.13 criterium; APA, 2000). This might implicate that children with PTSD are not able to report on their own feelings. However, this cannot explain the difference in findings between adults and children with regard to the RSDI and is therefore not the most likely explanation of the current findings.
Alternatively, it is also possible that the RSDI is too complicated for children. In the original factor structure (Hopper, Frewen, Sack et al., 2007), exceptionally high (> .90) standardized factor loadings were found. The high correlations that we found might indicate that there are some unnecessary items in the RSDI that make the RSDI too long or
complicated for children. It is also possible that the choice of words was too complex for (younger) children. Again, it would be interesting to compare younger children with older
children. If the RSDI turns out to be too complicated for children, a new questionnaire should be developed to assess state PTSD symptoms in children. To reevaluate and possibly develop a new questionnaire for state symptoms, the PTSD symptom data bank of Del Vecchio, Elwy, Smith, Bottonari and Eisen (2011) could be used. Del Vecchio and colleagues made a data bank of items from all available self-report PTSD measures, including items on
reexperiencing, avoidance and dissociation. These can be used in computer adapted testing, thereby possibly decreasing the number of questions people need to answer to get reliable results (Del Vecchio et al., 2011).
A fourth possible explanation for the different result in this study compared to the Hopper, Frewen, Sack et al., (2007) study is that in the current study no minimal PTSD severity score was required for the participants. In this study parent/caregiver scores were available and used to complement PTSD diagnosis, so children with very low self-reported trait symptom scores who did have a diagnosis of (partial) PTSD were used in this sample. These are children that either cannot report well on their own symptoms, or are very avoidant. The current sample is a representative clinical sample, therefore generalizable to the population of children with PTSD. In the Hopper, Frewen, Sack et al., (2007) study, only participants with high severity scores were used, which could mean that their sample is not representative of a clinical PTSD population because avoidant patients or patients that cannot report their own symptoms well are excluded. Unfortunately, the current sample was too small to look at children with higher PTSD severity scores only. Future studies could compare the factor structure of the RSDI in children who report low symptoms with children reporting higher symptoms. Furthermore, it would be interesting to study and the factor structure of the RSDI in a representative adult sample that also consists of people with lower severity scores.
It is not likely that the current results are caused by methodological concerns such as a small sample size. According to MacCallum, Widaman, Zhang and Hong (1999), a sample size smaller than 100 does not have to be a problem, if communalities are expected to be high. In the paper by Hopper, Frewen, Sack et al., (2007), communalities were high for the three factors in the RSDI. In the current study, communalities were also quite high, though not as high as in the original paper. However, it is a limitation of the current study that the sample size was too small to distinguish between age and severity groups.
In short, the RSDI cannot be used to measure state symptoms of PTSD in children aged 8-18 in the same way as in adults. This could imply a development in how PTSD manifests itself, with symptoms differentiating more with age. Alternatively, it is possible that children are not able to report well on their state PTSD symptoms, or that the RSDI is not the right tool to measure it. A strong point of this study is that a representative clinical group of children with PTSD was used, in contrast to earlier studies. Future research could shed light on this matter by comparing younger and older children and by studying a representative group of adult patients.
State and trait symptoms of PTSD, heart rate and heart rate variability
In line with our hypotheses, state measures of PTSD correlated with heart rate (HR) reactivity scores. Like in adults, more PSTD state symptoms were related to higher HR reactivity. However, because we analyzed two factors of the RSDI instead of the original three (Hopper, Frewen, Sack et al., 2007) these findings cannot be easily related to earlier studies with the RSDI and it is not possible to relate HR reactivity to different symptoms of PTSD. Since this is the first study to assess state symptoms of PTSD in children, this is a
finding that requires further research. It does show that state symptoms of PTSD correspond to heart rate, like in adults (e.g. Hopper, Frewen, Sack et al., 2007).
Unexpectedly, with regard to trait PSTD symptoms, no correlation with heart rate was found. Two concepts that are possibly important in the relationship that have not been taken into account here are PTSD severity and type of trauma. PTSD severity scores can moderate the responsivity of psychophysiological response to trauma-related cues (Orr & Roth, 2000). In the current study, severity scores were quite low when compared to other studies (e.g. Kirsch, Wilhelm & Goldbeck, 2015) and its possible influence was not controlled for. This might have influenced the results of the study negatively, because relatively few participants had high severity scores. However, the severity scores of the current sample represent a clinical sample of children with PTSD which means that they are all realistic values.
In addition to severity scores, type of trauma was not taken into account in this study. Two types of trauma are often distinguished, see for example Jonkman, Verlinden, Bolle, Boer and Lindauer (2013). Type I consists of a single traumatic experience (like a car accident) and type II of multiple or recurrent traumatic experiences (like child maltreatment). Jonkman et al., (2013) show that PTSD symptoms and trauma-unrelated symptoms are more severe in maltreated children compared to those who experienced a single trauma. In the current study, no distinction was made between the two groups. Most previous studies used a sample of children that experienced type I trauma (see for example the review by Kirsch, Wilhelm & Goldbeck, 2011). It would be interesting to see if children who experienced type II trauma show different responses than children who experienced type I trauma, but the current sample was too small to make such a distinction. Since type II trauma is associated with more severe PTSD symptoms than type I trauma (Jonkman et al., 2013), it is possible that the two different groups also differ on severity scores, thereby further complicating the interpretation of the
results. Future research could compare psychophysiological responses of children with type I and type II trauma, controlling for severity scores.
It is also possible that psychophysiological measures are not related to trait PTSD symptoms in children, in contrast to in adults. As described in the introduction, results with regard to psychophysiological studies have been heterogeneous in children (e.g. Jones-Alexander et al., 2005; Saltzman et al, 2005; Scheeringa et al, 2004). Kirsch, Wilhelm and Goldbeck (2015) found no relationship between heart rate and PTSD and they hypothesize that in children, higher heart rate is not associated with PTSD in the same way as in adults. They describe possible explanations for this difference, including higher resting heart rate in children compared to adults. It would be interesting to study a group of children and adults with comparable traumatic experiences in the same manner, to make a direct comparison between children’s and adults’ psychophysiological responses.
In this light, it is interesting that state symptoms do show a relationship with HR reactivity in the current study. Children rated their state symptoms about a trauma script during which their heart rate was also measured, whereas the trait symptoms of PTSD are rated over the time course of a month. This means that heart rate was measured at the same time as the state symptoms, perhaps leading to a larger correlation than with trait symptoms. Possibly, trait symptoms of PTSD are less stable in children than in adults, thus leading to a non-significant association with heart rate reactivity, while state symptoms show a more reliable image of current PTSD symptoms in children. This could imply that measuring state symptoms of PTSD is an important addition to diagnostical procedures that look at trait symptoms only. It would be interesting to replicate this finding in children, perhaps with another measure of state PTSD as discussed in the previous section of this study.
It would also be interesting to divide PTSD patients into different groups. So, instead of comparing the symptoms of all children to their heart rate, you could categorize children in groups according to their reported PTSD symptoms. PTSD is a very heterogeneous disorder: with criteria from the DSM-IV there are 79.794 possible combinations of symptoms that all qualify for the diagnosis PTSD (Galatzer-Levy & Bryant, 2013). Dividing PSTD patients by their symptom types might help resolving part of this heterogeneity. The groups can then be compared in their HR/HRV responses to trauma cues. If these groups differ in their
psychophysiological responses, this might cancel out results if you take all patients together. People whose PTSD symptom patterns show a reexperiencing type of PTSD might have very different psychophysiological responses to a trauma script than someone with an overall avoidant type of PTSD, even though their overall reexperiencing scores are similar.
Some characteristics of the current study make the findings reliable and interesting for future studies. First, PTSD trait symptoms have been assessed with a validated instrument in the current study. Second, the HR data shows a small but significant increase of the HR in the trauma compared to the neutral condition. This indicates a reaction of children to the trauma script comparable to that of adults. Also, HR reactivity correlates with state PTSD symptoms. Third, the current study is the first to investigate the relationship between heart rate and PTSD symptoms in children. Finally, the current sample is a representative clinical sample, making the results generalizable to the population of children with PTSD.
The current study shows no relationship between trait symptoms of PTSD and heart rate data. This is the first time this relationship has been studied in children and therefore it should be replicated before strong conclusions can be drawn about the absence of this association. However, the current sample consisted of a representative clinical group of children and the results are in line with some previous studies into heart rate and PTSD in
children. State symptoms were significantly associated with heart rate reactivity, comparable to in adults. It is yet unclear where the difference between trait and state symptoms in their connection to heart rate originates from, but possibly state measures are an important to take into account in diagnostic procedures to establish PTSD.
Literature
American Psychiatric Association. (2000). Diagnostic and statistic manual of mental
disorders (4th ed., text rev.). Washington, DC: Author.
Blake, D. D., Weathers, F. W., Nagy, L. M., Kaloupek, D. G., Gusman, F. D., Charney, D. S., & Keane, T. M. (1995). The development of a clinician-administered PTSD scale.
Journal of Traumatic Stress, 8, 75-90.
Buckley, T. C., & Kaloupek, D. G. (2001). A meta-analytic examination of basal
cardiovascular activity in posttraumatic stress disorder. Psychosomatic Medicine, 63, 585-594.
Contractor, A. A., Layne, C. M., Steinberg, A. M., Ostrowski, S. A., Ford, J. D., & Elhai, J. D. (2013). Do gender and age moderate the symptom structure of PTSD? Findings from a national clinical sample of children and adolescents. Psychiatry
Research, 210(3), 1056-1064.
Copeland, W. E., Keeler, G., Angold, A., & Costello, E. J. (2007). Traumatic events and posttraumatic stress in childhood. Archives of General Psychiatry, 64, 577-584. Deighton, J., Croudace, T., Fonagy, P., Brown, J., Patalay, P., & Wolpert, M. (2014).
Measuring mental health and wellbeing outcomes for children and adolescents to inform practice and policy: a review of child self-report measures. Child and
adolescent psychiatry and mental health, 8(1), 14.
Del Vecchio, N., Elwy, A. R., Smith, E., Bottonari, K. A., & Eisen, S. V. (2011). Enhancing self-report assessment of PTSD: Development of an item bank. Journal of Traumatic
Stress, 24(2), 191-199.
the Clinician Administered PTSD scale for Children and Adolescents in a Dutch Population. European Journal of Psychotraumatology, 4.
Fairbank, J. A., & Fairbank, D. W. (2009). Epidemiology of child traumatic stress. Current
Psychiatry Reports, 11, 289-295.
Galatzer-Levy, I. R., & Bryant, R. A. (2013). 636,120 Ways to have posttraumatic stress disorder. Perspectives on Psychological Science, 8(6), 651-662.
de Geus E. J. C., Willemsen, G. H. M., Klaver, C.H. A. M., & van Doornen L. J. P. (1995). Ambulatory measurement of respiratory sinus arrhythmia and respiration rate,
Biological Psychology, 41, 205-227.
Green, K. T., Dennis, P. A., Neal, L. C., Hobkirk, A. L., Hicks, T. A., Watkins, L. L., ... & Beckham, J. C. (2016). Exploring the relationship between posttraumatic stress disorder symptoms and momentary heart rate variability. Journal of Psychosomatic
Research, 82, 31-34.
Hopper, J. W., Frewen, P. A., Sack, M., Lanius, R. A., & van der Kolk, B. A. (2007). The responses to script-driven imagery scale (RSDI): Assessment of posttraumatic symptoms for psychobiological and treatment research. Journal of Psychopathology
and Behavioral Assessment, 29, 249-268.
Hopper, J. W., Frewen, P. A., van der Kolk, B. A., & Lanius, R. A. (2007). Neural correlates of reexperiencing, avoidance, and dissociation in PTSD: Symptom dimensions and emotion dysregulation in responses to script-driven trauma imagery. Journal of
Traumatic Stress, 20(5), 713-725.
Jones-Alexander, J., Blanchard, E. B., & Hickling, E. J. (2005). Psychophysiological
assessment of youthful motor vehicle accident survivors. Applied Psychophysiology
Jonkman, C. S., Schuengel, C., Lindeboom, R., Oosterman, M., Boer, F., & Lindauer, R. J. (2013). The effectiveness of Multidimensional Treatment Foster Care for Preschoolers (MTFC-P) for young children with severe behavioral disturbances: study protocol for a randomized controlled trial. Trials, 14(197).
Jonkman, C. S., Verlinden, E., Bolle, E. A., Boer, F., & Lindauer, R. J. L. (2013). Traumatic stress symptomatology after child maltreatment and single traumatic events: Different profiles. Journal of Traumatic Stress, 26, 225-232.
King, D. W., Leskin, G. A., King, L. A., & Weathers, F. W. (1998). Confirmatory factor analysis of the Clinician-Administered PTSD Scale: Evidence for the dimensionality of posttraumatic stress disorder. Psychological Assessment, 10, 90–96
Kirsch, V., Wilhelm, F. H., & Goldbeck, L. (2015). Psychophysiological characteristics of pediatric posttraumatic stress disorder during script-driven traumatic
imagery. European Journal of Psychotraumatology, 6.
Kirsch, V., Wilhelm, F. H., & Goldbeck, L. (2011). Psychophysiological characteristics of PTSD in children and adolescents: A review of the literature. Journal of Traumatic
Stress, 24(2), 146-154.
Lanius, R. A., Vermetten, E., Loewenstein, R. J., Brand, B., Schmahl, C., Bremner, D., & Spiegel, D. (2010). Emotion modulation in PTSD: Clinical and neurobiological evidence for a dissociative subtype. American Journal of Psychiatry, 167(6), 640-647. MacCallum, R. C., Widaman, K. F., Zhang, S., & Hong, S. (1999). Sample size in factor
analysis. Psychological Methods, 4(1), 84-99.
(2008). The posttraumatic stress disorder diagnosis in preschool-and elementary school-age children exposed to motor vehicle accidents. American Journal of
Psychiatry, 165, 1326-1337.
Nader, K.O., Newman, E., Weathers, F.W., Kaloupek, D.G., Kriegler, J.A., & Blake, D.D. (2004). National Centre for PTSD Clinician-Administered PTSD Scale for Children
and Adolescents (CAPS-CA) Interview Booklet. Los Angeles: Western Psychological
Services.
Orr, S. P., & Roth, W. T., (2000). Psychophysiological assessment: clinical applications for PTSD. Journal of Affective Disorder, 61(3), 225-240.
O’Connor, B. P., (2000). SPSS and SAS programs for determining the number of components using parallel analysis and Velicer’s MAP test. Behavior Research Methods,
Instruments & Computers, 32, 396-402.
Pitman, R. K., Orr, S. P., Forgue, D. F., Altman, B., de Jong, J. B., & Herz, L. R. (1990). Psychophysiologic response to combat imagery of Vietnam veterans with posttraumatic stress disorder versus other anxiety disorders. Journal of Abnormal
Psychology, 99(1), 49-54.
Pole, N. (2007). The psychophysiology of posttraumatic stress disorder: a meta-analysis.
Psychological Bulletin, 133(5), 725-746.
Saltzman, K. M., Holden, G. W., & Holahan, C. J. (2005). The psychobiology of children exposed to marital violence. Journal of Clinical Child & Adolescent Psychology,
34(1), 129-139.
Scheeringa, M. S., Zeanah, C. H., Myers, L., & Putnam, F. (2004). Heart period and
variability findings in preschool children with posttraumatic stress symptoms.
Silverman, W. K., & Albano, A. M. (1996). The anxiety disorder interview schedule for
DSM-IV: Child interview schedule. San Antonio, TX: Graywind.
Spielberger, C. D. (1973). Manual for the state-trait anxiety inventory for children. Palo Alto, CA: Consulting Psychologists Press.
Stein, M. B., Walker, J. R., Hazen, A. L., & Forde, D. R. (1997). Full and partial
posttraumatic stress disorder: findings from a community survey. American journal of
psychiatry, 154(8), 1114-1119.
Tabachnick, B. G., and Fidell, L. S. (2013). Using Multivariate Statistics , 6th ed. Boston : Pearson.
Tan, G., Dao, T. K., Farmer, L., Sutherland, R. J., & Gevirtz, R. (2011). Heart rate variability (HRV) and posttraumatic stress disorder (PTSD): A pilot study. Applied
psychophysiology and biofeedback, 36(1), 27-35.
Verlinden, E., Schippers, M., van Meijel, E., Beer, R., Opmeer, B. C., Olff, M., Boer, F., & Lindauer, R. J. L. (2013). What makes a life event traumatic for a child? The predictive vales of DSM-criteria A1 and A2. European Journal of
Psychotraumatology, 4.
Weathers, F. W., Ruscio, A. M., & Keane, T. M. (1999). Psychometric properties of nine scoring rules for the Clinician Administered Posttraumatic Disorder Scale.
Psychological Assessment, 11, 124-133.
Wolf, E. J., Miller, M. W., Reardon, A. F., Ryabchenko, K. A., Castillo, D., & Freund, R. (2012). A latent class analysis of dissociation and posttraumatic stress disorder: Evidence for a dissociative subtype. Archives of General Psychiatry, 69(7), 698-705.
Table 1
Pearson Correlations, Means and Standard Deviations of the scores on the RSDI
Item 1 2 3 4 5 6 7 8 9 10 11 1 1 2 .366** 1 3 .251* .468*** 1 4 .593*** .425*** .393** 1 5 .665** .249* .235* .717*** 1 6 .533** .352** .474*** .762*** .673*** 1 7 .339** .241* .360** .505*** .393** .495*** 1 8 .558*** .234* .356** .583*** .551*** .467*** .454*** 1 9 .568*** .414*** .299* .483*** .580*** .454*** .520*** .710*** 1 10 .183 .356** .262* .235* 0.054 0.185 0.056 -0.02 0.033 1 11 .315** .664*** .432*** .379** 0.225 .452*** 0.208 .236* .334** .301* 1 M 4.15 3.07 2.51 3.35 3.63 3.57 3.39 3.71 3.92 2.71 3.29 SD 1.83 2.08 2.04 2.02 2.11 2.15 2.05 1.97 1.96 2.23 2.27 Note. * p < .05, ** p < .01, *** p < .001
Table 2
Factor pattern, factor structure matrices and communalities for the EFA
RSDI item
Factor pattern matrix Factor structure matrix Communalities Factor 1 Factor 2 Factor 1 Factor 2 Initial Extraction
1 .693 .073 .724 .371 .549 .529 2 .077 .783 .414 .816 .593 .670 3 .231 .467 .432 .566 .410 .364 4 .755 .157 .822 .482 .746 .696 5 .880 -.139 .820 .239 .693 .688 6 .675 .204 .762 .494 .699 .615 7 .576 .031 .589 .279 .412 .348 8 .801 -.114 .752 .231 .640 .576 9 .736 .019 .744 .335 .677 .553 10 -.086 .487 .123 .450 .220 .209 11 .072 .736 .389 .767 .520 .593
Note. Factor structure coefficients larger than |.40| are retained for that factor and are shown in italics.
Table 3
Number, minimum, maximum, means and standard deviations for CAPS scales, RSDI factors and heart rate measures.
N Minimum Maximum M SD Reexperiencing 69 0 35 16.41 9.82 Avoidance 69 0 16 7.54 3.83 Dissociation 58 0 15 3.14 3.77 RSDI factor 1 72 2 54 31.24 13.13 RSDI factor 2 72 0 36 18.50 9.11 HR neutral 74 46.40 112.87 79.59 13.02 HRV neutral 72 14.50 327.00 109.58 63.68 Log HRV neutral 74 1.16 2.51 1.97 0.27 HR trauma 73 49.51 117.99 81.29 14.13 HRV trauma 73 8.00 265.00 101.01 55.11 Log HRV trauma 74 0.90 2.42 1.93 0.30
Table 4
Correlations between CAPS scales and HR and HRV reactivity scores, r (N)
Reexperiencing Avoidance Dissociation RSDI factor 1 RSDI factor 2 HR reactivity HRV reactivity Reexperiencing 1 Avoidance .455** (69) 1 Dissociation .506** (58) .409** (58) 1 RSDI factor 1 .616** (69) .423** (69) .338** (58) 1 RSDI factor 2 .548** (69) .460** (69) .250 (58) .815** (72) 1 HR reactivity .189 (67) .184 (67) -.105 (57) .278* (70) .243* (70) 1 HRV reactivity -.043 (66) -.086 (66) .053 (56) -.080 (69) .058 (69) -.400** (72) 1 Note.* p < .05, ** p < .01
